Capacitive transducers with a flexible sensing diaphragm convert an applied force, pressure, or physical displacement to a change in capacitance. This capacitive change is transduced by an electrical circuit to a corresponding change in electrical voltage, current, or frequency. Prior art capacitive transducers use substantially parallel-plate electrodes separated a small distance apart in vacuum or a fluid dielectric medium. The sensitivity, linearity, and dynamic range of capacitive transducers is limited by the disadvantages of such variable parallel-plate capacitors.
When fringe fields are ignored, the capacitance C between two, conducting parallel plates is substantially given by: EQU C=.di-elect cons.A/d,
where .di-elect cons.is the permittivity of the dielectric medium, A is the effective area of the capacitor plates, and d is the effective spacing between the capacitor plates. Capacitance-displacement sensitivity, the change in capacitance as a function of plate spacing, is given by: EQU .DELTA.C/.DELTA.d=-.di-elect cons.A/d.sup.2
which has a dependency on d.sup.2 which results in a non-linear increase in capacitance sensitivity with decreasing plate spacing.
The capacitance-displacement sensitivity of a transducer with substantially parallel plates also can be affected by:
1. the non-linear reduction in capacitive sensitivity due to bending stresses in the sensing plate when the ratio of plate deflection to plate thickness is substantially greater than 0.2;
2. the non-linear reduction in capacitive sensitivity due to tensile stresses arising from the stretching of a thin sensing plate or diaphragm; and,
3 the reduction in capacitive sensitivity and frequency response due to viscous damping when a fluid dielectric, such as air, is squeezed between the capacitor plates.
U.S. Pat. No. 5,048,165 issued Sep. 17, 1991, discloses a method to construct a capacitive transducer with a deformable plate located between two, rigid plates. Differential capacitance detection allows two different and oppositely-sensed non-linearities to cancel to extend the linear range of the transducer. The disadvantage of this method is the complexity of using an additional capacitor plate and the requirement to construct a three-plate capacitive structure with well known, closely maintained and matched mechanical, thermal, and electrical characteristics.
U.S. Pat. No. 4,996,627 issued Feb. 26, 1991, discloses a three-plate, capacitance transducer used with an electronic circuit disclosed in U.S. Pat. No. 5,019,783 issued May 28, 1991, to provide a linear electrical output for a transducer with intrinsic non-linear sensitivity. U.S. Pat. 4,584,885 issued Apr. 29, 1986, discloses another of the many electronic circuits devised to electrically linearize the outputs of capacitive transducers. The disadvantages of these approaches is the requirement to use a third capacitor plate and the complexity and cost of signal compensation electronics. A general disadvantage of using mechanical or electrical methods to linearize the response of a capacitive transducer with substantially parallel-plate electrodes is that the sensitivity and dynamic range of the transducer cannot be significantly increased.
Other disadvantages and limitations of prior-art capacitive transducers arise from low values of quiescent capacitance. The maximum quiescent capacitance of a capacitive transducer is determined by the minimum spacing that can be reliably maintained between parallel-plate electrodes. Plate spacing is limited by the dimensional tolerances and stability of precision components and support structure. Plate spacing can also be limited by the voltage applied across the electrodes.
For microphones and capacitance transducers with thin sensing diaphragms, electrode spacing is further restricted by the space required to accommodate diaphragm displacement. As an example, a Bruel & Kjaer Model 41444, one-inch diameter, research-grade, capacitive microphone maintains a nominal 20-micron gap between a thin, nickel diaphragm and a rigid back-plate electrode. This spacing limits microphone capacitance to typically 55 pF and requires the device to be constructed from thermally stable components with precision tolerances. A 20-micron, dielectric gap is 100 to 1000 times larger that the thickness of dielectric films, such as silicon dioxide and silicon nitride, that are used to construct integrated circuit devices.
Low values of quiescent capacitance C.sub.o in capacitive transducers can cause a loss in sensitivity due to parallel stray capacitance. The total stray capacitance C.sub.s of support structure, electrodes, conducting leads, and the inputs of electronic circuits that shunts the quiescent capacitance reduces sensitivity by a factor C.sub.o /(C.sub.o +C.sub.s). Stray capacitance is of particular concern for transducers constructed with small, micromachined components and thin material layers. Low-capacitance transducers are more susceptible to electromagnetic interference and to changes in stray capacitance compared to transducers with higher quiescent capacitance.
Another disadvantage of transducers with parallel-plate capacitors is the increased noise in electrical networks with small capacitors. It is well known that the mean-squared voltage fluctuation, .DELTA.V.sup.2 of a system with a capacitor at thermal equilibrium equals kT/C where k is Boltzmann's constant and T is absolute temperature. This noise source limits the accuracy and dynamic range of a capacitive transducer when it exceeds fundamental noise resulting from thermally induced motion of the sensing diaphragm.
Yet another disadvantage of a parallel-plate capacitive transducer, with a thin sensing diaphragm, is the maximum voltage that can be safely applied across the capacitor plates. Large displacements of a thin diaphragm resulting from shock or over-pressure loads can cause the diaphragm to collapse against its counter electrode. This occurs when the diaphragm deflects to a position where electrostatic force overcomes the mechanical restoring force of the diaphragm.
Capacitive transducers used to measure acceleration frequently use electrostatic force-feedback to maintain a suspended proof mass in a substantially fixed location. This minimizes non-linear capacitance sensitivity with electrode spacing. However, feedback cannot increase capacitance sensitivity or overcome the disadvantages of small quiescent capacitance limited by practical electrode spacing.
A variable capacitor has linear response if the area of the capacitor plates are changed while the plate spacing remains fixed. This can be accomplished by moving or rotating multiple plates in parallel planes. This approach was used to capacitively tune early radios, but is difficult to implement in small transducers.
An article titled "A capacitor transducer using a thin dielectric and variable-area electrode" appearing in the IEE Proc., Vol. 127, Pt. A, No. 6, July 1980, by Basarab-Horwath et al., reports high values of capacitive sensitivity for a transducer with capacitor plates that increase in area with applied force. The disadvantage of this transducer is that the shape of the flexible electrode changes with both displacement and applied loading. Therefore, it is difficult to obtain, maintain, and control a precision capacitance relationship between the electrodes. This article does not teach or suggest the benefits of using a rigid electrode with a surface contour chosen to obtain an accurate, repeatable, and specific capacitive relationship between the electrodes of a variable capacitor or a capacitance transducer.
The work by Basarab-Horwath et. al. is reported as an extension of the work by Caterer et al. described in "Measurement of Displacement and Strain by Capacity Methods", Proc. J. Mech. E., (152) 1945. Carter et al. describe a variable capacitor with a tangential strip electrode that deforms in an arc around an electrode of cylindrical cross-section. This article does not teach or suggest the benefits of using the larger surface perimeter of a diaphragm or plate to obtain higher values of capacitance sensitivity or how to linearize capacitance sensitivity for flexible electrodes with different deflection-load response characteristics. This type of variable capacitor also has the disadvantage of not having integral, self-supporting components in a compact and rigid assembly.
U.S. Pat. No. 4,225,755, issued Sep. 30, 1980, discloses two embodiments of a capacitive force transducer primarily for use as a microphone. A first embodiment comprises a thin metal diaphragm held in contact to a dielectric film on a metal electrode having a lip of cylindrical cross-section. A second embodiment comprises a thin conducting diaphragm suspended over dielectric material adhered to an electrode that is anisotropically etched from crystalline material. An advantage attributed to the two embodiments is small quiescent capacitance. The capacitance of the second embodiment is too small to be practically utilized in a transducer as it is severely limited by the large angle that exits between the &lt;100&gt; and &lt;111&gt; planes of a crystalline material with cubic diamond crystal structure. This invention does not teach or suggest the advantages of electrodes contoured to provide high values of quiescent capacitance and correspondingly high values of capacitance sensitivity. This patent and an associated patent, U.S. Pat. No.4,360,955, issued Nov.30, 1982, also do not teach or suggest the benefits of using a rigid electrode with a specific surface contour to control capacitive sensitivity with diaphragm deflection, to maximize the linear dynamic range of the transducer, or to linearize capacitive sensitivity for flexible electrodes with different deflection-load response characteristics.
Accordingly, the present invention was developed to provide a capacitor transducer with high capacitive sensitivity that is governed by the rate of change of the effective area of the capacitor electrodes; that is independent of electrode spacing; and that has the advantages of the narrow electrode spacing provided by a thin film dielectric spacer.
Many electronic circuits have been devised to transduce the changes of a variable capacitor. The most sensitive and stable circuits utilize ratiometric bridge networks, but none are known to have the low-noise, high sensitivity, and wide linear dynamic range to fully exploit the performance capabilities of the variable capacitor of the present invention. This capability can be realized if a capacitive detection circuit can be devised that has in combination the following features and performance capabilities:
a. a circuit arrangement with a ratiometric bridge-like network to detect small differences in the capacitance between a variable capacitor and a stable reference capacitor; PA1 b. a circuit in which a transconductance amplifier feeds back current to null a bridge network for large capacitance changes. PA1 c. a feedback circuit with a linear output voltage which allows one electrode of a variable capacitor to be grounded; PA1 d. a feedback circuit that is stable at low values of closed-loop gain to accommodate large capacitive changes up to 1000% and more; PA1 e. a circuit having a wide linear dynamic range of 120 dB, and more, at low frequencies down to DC; PA1 f. a low-impedance, circuit arrangement that minimizes thermal noise of passive components and the voltage and current noise of an amplifier used for closed-loop bridge balancing; PA1 g. a circuit with differential loss-pass filtering before amplification; PA1 h. a circuit with an amplifier for which the input impedance, dynamic response, bandwidth, and common-mode rejection of the inverting and non-inverting inputs are substantially identical; PA1 i. a bridge-like network that minimizes signal division by fixed elements and uses the majority of the time during an excitation cycle to develop a measurement signal; PA1 j. a circuit that can to operate from a single, low-voltage power supply or from higher-voltage, bipolar supplies; PA1 k. a circuit for which active shielding can be easily implemented. PA1 a. an intrinsically linear output that does not require a third electrode or additional electronic circuitry to linearize capacitance sensitivity. PA1 b. high values of active quiescent capacitance to reduce losses in sensitivity and dynamic range due to stray capacitances and electrical noise. High values of quiescent capacitance also reduce transducer susceptibility to electromagnetic interference and changes in parasitic and stray capacitance. PA1 c. a thin dielectric layer such as silicon dioxide or silicon nitride that can be reliably grown or deposited in substantially 20 to 200 nm thick layers by well established integrated circuit manufacturing methods. PA1 d. an electrode configuration for which capacitance sensitivity and frequency response are not dampened by a fluid between the electrodes such as air. PA1 e. a thin sensing diaphragm that is not vulnerable to electrostatic collapse. PA1 f. an open electrode configuration that accommodates large deflections of the flexible electrode or deflections of a diaphragm with a integral hub. PA1 g. a rigid electrode with a central aperture to allow fluid pressure to be applied to the backside of a sensing plate or diaphragm to allow measurements of differential pressure and fluid flow. PA1 h. a flexible electrode fabricated from single-crystal silicon. Silicon is substantially free of hysteresis because of its extraordinary elastic properties and silicon diaphragms can be fabricated from silicon wafers with thicknesses down to one micron, or less. Thickness uniformity and the control of internal stress in silicon diaphragms is superior to metal foils. PA1 a. A ratiometric bridge-like network that detects small differences in capacitance between a variable sensing capacitor and a fixed, stable reference capacitor, or alternatively between two variable capacitors. The use of a bridge minimizes errors associated with phase and timing variations of the bridge excitation waveform as well as errors arising from common-mode electromagnetic interference. PA1 b. A circuit with a transconductance amplifier, or a voltage-controlled current source, to feed back current to null a ratiometric bridge network for changes in capacitance. The voltage used to control the feedback current being substantially linear with capacitance changes .DELTA.C/C up to 1000% and higher; PA1 c. A feedback circuit arrangement with a linear output that allows one electrode of a variable capacitor to be grounded, or alternatively one electrode of two variable capacitors to be grounded. Grounded electrodes minimize parasitic and stray capacitance that cause signal loss and electrical noise from capacitively coupled electrical fields. Grounding the variable capacitor element also eliminates the requirement for a separate signal return line between a capacitive transducer and its associated electronics. PA1 d. A circuit with a stable output at low values of closed-loop gain that allows the detection of the very large changes of the variable capacitor. PA1 e. A circuit with a low-noise output over a bandwidth from DC to the highest frequency at which capacitance changes are required to be detected. PA1 f. A circuit with a bridge-like network that has two resistance arms that are lower in impedance than conventional capacitance bridge circuits with four capacitors. This reduces thermal noise and allows an operational amplifier to be selected that has a combined value of voltage noise and low-frequency flicker (1/f) voltage noise that is comparable to effective values of its current noise and low-frequency flicker (1/f) current noise. The use of low-impedance circuitry also reduces circuit susceptibility to electrical pickup and minimizes signal loss due to voltage division across stray capacitances. A another advantage of low-impedance circuitry is it limits low-frequency drift associated with thermally induced changes in bias currents in differential amplifiers. PA1 g. A circuit in which the bridge-excitation frequency can be increased to megahertz levels to further reduce the impedance of the capacitive detection circuitry and the values of its bridge resistors. PA1 h. A circuit with a differential transconductance amplifier for which the input impedance, dynamic response, bandwidth, and common mode rejection of the inverting and non-inverting inputs are substantially identical. A differential transconductance amplifier with active feedback also has the advantage of accommodating large differential and common mode input signals. PA1 i. A circuit with low-pass filtering ahead of differential amplification to substantially reduce the fundamental and higher order frequencies of the bridge excitation voltages. This allows amplification and closed-loop control to be performed at only the highest required detection frequency, where amplifier gain and common-mode signal rejection are high compared to the bridge excitation frequency. Differential passive filtering also provides a substantially constant voltage to discharge the variable capacitor during a single, charge-discharge cycle. PA1 j. A detection circuit that maximizes signals across variable capacitors and minimizes signal division across fixed bridge components, and one that uses the majority of the time during an excitation cycle to develop a differential capacitance measurement signal. The capacitive sensitivity of the capacitance detection circuit of the present invention, in terms of the change in output voltage .DELTA.V for a given change in capacitance .DELTA.C/C is comparable or higher than that of prior art bridge circuits. Since the bridge capacitors are charged to the peak potential of a short excitation pulse, a loss of one-half does not occur due to voltage division across substantially equal capacitors in adjacent arms of a bridge network. Also, the sensed capacitor and the reference capacitor are simultaneously charged during the a period of time t.sub.1 that is short compared to the discharge time t.sub.2. This increases the average value of the differential bridge output over repetitive measurement cycles. PA1 k. A circuit that can operate from a single low-voltage power supply as well as from higher-voltage bipolar supplies to accommodate a broad range of capacitive-based measurement applications. PA1 l. A circuit arrangement for which active shielding can be used to minimize noise and electrical pickup from stray electrical fields and to minimize signal loss across stray capacitance.
Prior art capacitive detection circuits do not have a combination of all the above advantages. For example, circuits that use voltage feedback to achieve a linear response generally do not have low-impedance circuitry or allow one electrode of a variable capacitor to be grounded. Low-impedance circuits have a linear response over a very limited range.
Accordingly, the instant invention was developed to provide a capacitance detection circuit with the above features and capabilities to enhance the sensitivity, accuracy, and dynamic range of many different types of capacitance transducers, and specifically to fully exploit the performance capability of transducers using a variable capacitor of the present invention.